EU Decision-making and the Allocation of Responsibility

Manfred J. Holler[*]

First version: February 3, 2011

Revised: May 10, 2011

Prepared for the

Research Handbook on the Economics of European Union Law.

To be published by Edward Elgar, Cheltenham

Abstract: The underlying hypothesis of this paper is that the allocation of responsibility will improve decision making just like the assignment of property rights improves the allocation of goods. Similar problems exist in the case of collective decision making as in the provision of public goods. The paper contains some key results of applying the voting power approach to European Parliament, the Council of Ministers, and codecision making involving both institutions. The close relationship between freedom of choice and voting power is elaborated using formal models to represent both concepts. The focus will be on the Public Good Index of power. With the help of a similar model, the close relationship between power and causality can be demonstrated. This is shown by focusing on the NESS concept of causality and applying it to responsibility. A discussion of the various problems embedded in the power-responsibility hypothesis within the context of EU decision making concludes the paper.

Keywords: EU decision making, collective choice, political responsibility, causation, voting power

1. On political responsibility

When on March 25th, 1957, the Treaty of Rome was signed creating the European Economic Community (EEC) of The Six, the legislative and executive power was installed in the Council of Ministers – the Council in what follows. The voting weights were 4 votes each for the larger countries France, Italy, and West Germany, 2 votes each for Belgian and the Netherlands, and 1 vote for Luxembourg. Given the differences in size of the various countries, irrespective of whether measured by population or total GDP, it seemed that Luxembourg was highly overrepresented. However, a qualified majority rule of 12 votes was prescribed and, given this quota, Luxembourg had no say in cases that were submitted to a formal vote. For the given vote distribution w = (4, 4, 4, 2, 2, 1) and the decision rule d = 12, its a priori voting power was zero.

Can we conclude that the citizens of Luxembourg were not responsible for what happened in the EEC? Perhaps not, because in practical terms decisions by the Council relied on a full consensus, and the unanimity rule assigns to Luxembourg the same amount of voting power as it assigns to France, Italy or West Germany. However, could Luxembourg afford to say no if the five other EEC members said yes? Hardly![1] Does this imply that the government and the citizens of Luxembourg were not responsible for what happened in the EEC? Who is responsible for EU decision making: The Council, the Commission, the European Parliament (EP), the voters who selected the member of the EP or the voters who selected the members of their national governments, directly or via a parliament, the members of the national parliaments and their party peers who decide on government coalitions, the national governments that are represented in the Council, or the politicians, voters and bureaucrats who, in the past, designed the rules and defined the objects of today’s decision making in the EU? What follows if we can answer this question and allocate responsibility to particular institutions and their members? What is responsibility?

The model underlying the following analysis starts with a simple relationship between actions (A) and outcomes (O). In our context actions A are the result of collective decision making and outcomes O can be public, merit, or private goods - however, our focus will be on public goods. We identify A with decision making, a perhaps heroic assumption, given bureaucratic impediments, on the one hand, and the sometimes overly optimistic view on feasibility, on the other. The inputs in decision making are power (P), i.e. the potential to act, and the freedom of choice (F), i.e., the set of alternatives that is available to choose from. Responsibility is embedded in the evaluation (E) of O and the possibility to sanction (S), honor or punish, A and therefore the choice and decision making that resulted in A. Thus, E and S represent responsibility in our model. (The diagram in the appendix illustrates this model.)

This model and its analysis is motivated by the expectation (or hope) that the allocation of responsibility to individual decision makers will improve choices. This necessitates to clarify the causation that links actions A to outcomes O and also to relate causality to individual decision makers. The latter is solved by linking decision making to power P and the freedom of choice F, both serve as a proxy of free will that completes the responsibility model in the case of individual decision making. The hypothesis underlying this paper is that under a huge pile of provisos we can specify the causal impact of an individual decision maker in a collective decision process by P and thus measure his or her responsibility by P and E. The expectation is that the allocation of responsibility will improve decision making just like the assignment of property rights can improve the allocation of goods. However, quite similar problems exist in the case of collective decision making as in the provision of public goods.

The discussion and conceptualization of these problems with reference to EU decision making is the core of this paper. Section 3 presents some key results of applying the voting power approach to European Parliament, the Council of Ministers, and codecision making involving both institutions. In section 4, the close relationship between freedom of choice and voting power is elaborated using formal models to represent both concepts. The focus will be on the Public Good Index of power. With the help of a similar model the close relationship between power and causality can be demonstrated. This is shown, with a focus on the NESS concept of causality, in section 5. These sections offer material that should be of interest also from a methodological point of view. Section 6 will integrate these building blocks and relate them to reponsibility. The next section, i.e., section 2, contains an overview of the various problems embedded in the power-responsibility hypothesis with focus on EU decision making and a further illustration of the responsibility model presented in this paper. This section can be read as an extended introduction or as a summary note.

In the sequel the reader will repeatedly find references to power measures, more specifically, to the Banzhaf index, the Shapley-Shubik index and the Public Good Index (PGI). This is not the occasion to discuss these measures in detail, but a short introduction could be of help to those readers who have had as yet no experience with these measures. All three are based on the concept of a swing (or critical) player, i.e., a player that can turn a winning coalition into a losing one and a losing coalition into a winning one. In its normalized form, the Banzhaf index of player i, i, is equal to the number of coalitions that have i as swing player divided by the total number of swing positions in the (voting) game, summed over all players (voters). The PGI differs from the Banzhaf index inasmuch as only minimum winning coalitions (MWC) are considered. S is a MWC coalition if S\{i} is a losing one, for all i  S, i.e., all players of a MWC coalition have a swing position. The PGI of player i, labelled hi, counts in how many MWC i is a member and divides this sum by the sum of all swing positions the players have in the all MWC of the game. The Shapley-Shubik index is slightly more complicated as it takes the orderings of players in a coalition into consideration. The swing player metamorphoses into a pivot that turns a losing coalition into a winning one, taking the sequence of coalition members into account. The index value for player i counts the number of pivot positions of i in the voting game under consideration and divides this number by the number of all permutations that can be formed out of the n players of the game, i.e., n!. Below you find some simple examples that illustrate the Banzhaf index and the PGI and their application. Applications of the Shapley-Shubik index are also discussed. For those readers who want to learn more about these measures a substantial amount of literature is available that deals with power measures. Felsenthal and Machover (1998) is definitively the most substantial text in this area. However, in what follows we will not presuppose an intimate familiarity with these measures.

2. Responsibility, causation and power

Definitions of responsibility often include a moral component referring to duties, guilt, blameworthiness, or even shame. This is especially the case when it comes to collective responsibility.[2] So far the EU is hardly an institution that arouses moral sentiments despite its decision making on immigration, environmental standards, and security policy, areas in which moral norms seem to be highly relevant. In general, however, questions of efficiency are brought forward when it comes to EU decision making either defined as straightforward economic efficiency and international competitiveness or combined with the more sophisticated problem of balancing the preferences of the citizens of its member states.

The latter is an issue of preference aggregation and will be only briefly discussed in this paper. The responsibility notion of this paper focuses on an evaluation of decision making and its outcome. It implies that sanctions or, more generally, accountability will be demanded if the evaluation indicates poor performance. If there is a causal relationship between decision and outcome, then the assignment of responsibility seems to be straightforward and sanctioning only an issue of power. If there is a power to sanction, then performance can be expected to improve or to be even optimal (in some sense). This is the case of liability and tort law. Both assume causal relationships. However, EU decisions are collective choices. The choices may still be evaluated by their outcomes. However, the allocation of responsibility to individual decision makers and the assignment of responsibility and accountability (e.g. sanctions) is not obvious as there is no straightforward causality that links the outcome of collective decision making to the individual decision maker. It is not clear how responsibility can improve the decision making process without solving these problems.

A first approach to an appropriate solution suggests allocating responsibility in accordance to the power of the individual decision entity (see Holler 2007). In the case of the Council, such an entity is the national government that represents a particular member state. Below we will discuss several studies that measure the power of the national representation in the Council. The results, the power distributions in the Council, could serve as a first approach to allocating contributions to the EU budget or CO2 emission rights within the Union.

The agents who select these governments, i.e., voters, are secondary decision makers. Their impact on the outcome of the Council decisions is much more difficult to evaluate and so far only very stylized designs have been studied, e.g., the Penrose square root rule that is supposed to guarantee the equal a priori voting power of every EU citizen in the decision making of the Council. One of the problems with the Penrose square root rule is that it does not guarantee that citizens have equal power when it comes to selecting their national governments and thus may not ensure the representation of their preferences in the Council. The British first-past-the post system does not even guarantee that a clear majority of votes will lead to a majority of seats in the Westminster Parliament. On the other hand, proportional representation often necessitates coalition formation and a coalitional compromise in policy making, which is also the case in EU decisions. Voters who voted for party A, and against the competing party B, often are frustrated by finding A and B in the same government coalition after the election. In such cases it is rather difficult to trace the link between the voters’ preferences and the political outcome.

In general, on the EU level, the power structure is even more complex than on the national level as the EP and the Commission also have a say in decision making, and EP members seem to be cross-pressured by national and ideological interests, the latter deriving from their party affiliations. In what follows I will consider this relationship in a very stylized way without going into institutional and legal details that underlie the structure of this relationship. Tsakatika (2008) identifies three institutional requirements for political responsibility: accountability, openness and identifiability. He concludes that in the EU there is a “responsibility deficit” across all three dimensions but especially concerning accountability and identifiabiliy. The latter, he argues, is a result of the complexity of power relations in EU decision making which hinders the identification of the agent responsible for a given political outcome. Interestingly, this conclusion relates power and responsibility – two concepts that are not always seen in close relationship with each other. Thus, it seems to be an adequate point of departure to study the power structure of this organization. In what follows, some tools will discussed and illustrated which should help to create a legal framework such that political or economic responsibility works.

In a democratic system, political responsibility functions to legitimate rulings through voting and to sanction poor performance by voting for an alternative. This however can only work if the legal system allows for political competition. Elections (and voting) are the standard tools to implement political competition. However, countervailing power, the control of one decision making institution by another and the need for a bargaining solution, is another mechanism that implements competition and allows for political responsibility. This is about checks and balances. In a democratic society this mechanism involves representation and, again, voting as a means of selection but also of sanctioning. Liability derives from legal responsibility and accountability presumes a straightforward causal relation. Both do not directly apply to EU decision making as it concerns the legislative institution itself and as it involves collective choices.

EU policy, also due to its historical background, has a strong impact on the economic performance and the structuring of the private economic sector. With reference to EU standardization politics, I will argue below that the EU tends to shift duties and solutions to the private sector when political responsibility becomes a burden. Political responsibility is traded with economic responsibility. The idea of economic responsibility by and large coincides with the invisible hand: economically inefficient projects get punished by deficit and perhaps bankruptcy, while efficient projects are profitable, successful and will multiply as long as there is demand. The market mechanism takes over responsibility. But where are the sanctions to the designers of the market if market performance is poor? As noted by Adam Smith, self interest and the invisible hand only “frequently” promote the interest of the society (Smith 1981 [1776/77]: 456),

In Goerke and Holler (1998),[3] we argued that the New Approach to European Standardization (NAES), laid out in the Commission’s Green Paper of October 1990, was a means of the EU authority to get rid of the responsibility of European standardization, a project that looked unprofitable from a political point of view. NAES stipulates that, in general, standards will no longer be decided upon by the Commission in conjunction with the Council. Instead, the Commission is meant to establish standards (a) with the help of European standardization bodies, (b) via providing direct or indirect incentives to companies which should apply standards and should contribute to the financing of the standardization process, and (c) with the support of the national standardization bodies. The mutual recognition (MR) of national regulations on product specifications and of decisions of the European Court of Justice (ECJ) concerning the goods market can be considered as an implication of NAES.

It could be argued that by renouncing the Old Approach that relied on detailed harmonization, the Commission has shed its influence on standardization in Europe to a large degree and has handed over its responsibilities to fairly independent bodies such as the CEN (Comité Européen de Normalisation), CENELEC (Comité Européen de Normalisation Electrotechnique) and ETSI (European Telecommunications Standards Institute).[4] The introduction of MR is a further step in this direction. In Goerke and Holler (1998), taking a public choice perspective, we asked why the Commission should have engineered this deprivation of formal power in European standardization? Why does the Commission rely on the interaction of nominally independent agents in organizing European standardization when it seems a relevant hypothesis to assume that these agents act, at least predominantly, in their own interest or in the interest of those who they are funded by? Is MR an adequate instrument to bring the internal goods market closer to efficiency despite the self-interested agents in the standardization game? Does regulatory competition work the same as competition on goods markets is supposed to work? Does it enhance efficiency?

Or is NAES just an instrument of the EU authorities, the Commission and the Council, to dispose of the responsibility for European standardization - and also to avoid explicit conflicts due to vested interests? If so, then the introduction of MR can be viewed as a means of obfuscation in this policy arena. Government politicians could be at a disadvantage in popularity and lose votes on the national level if they are held responsible for the standardization policy of the Commission and the Council, especially if the resulting harmonization is felt tremendously inflexible and biased in favor of some (interest-) groups in the population. The Old Approach, based on detailed (technical) standardization, was not very popular, sometimes even subject to satire and jokes, and it was rather costly. In comparison, it appears that MR obfuscation policy works properly. On the one hand, it seems to give back some regulatory power to national institutions, on the other, it seems not only to favor the market of regulations but also support competition between the regulators and even in the internal goods market. But these effects are far from obvious. For instance, in a recent publication, Kerber and van den Bergh (2008) argued that MR leads to a number of inconsistencies: Instead of preserving decentralized regulatory powers and supporting regulatory competition, MR is primarily a path to convergence and rather rigid harmonization. On the other hand, Pelkmans (2011: 22 of ms) concludes: “Mutual recognition is a great invention of the EU.” Regulatory MR, “has been very successful over time…The combination of an even more effective approach to existing barriers and an intrusive and targeted pre-emption policy for future ones has effectively spared the single goods market from destructive erosion.” In Pelkmans (2007: 699), he observes that MR “is rightly applauded as an ingenious innovation by economists, lawyers and political scientists alike.” But Pelkmans also talks of disillusions. He points out that